Back
[00:04:18] -!-
theorbtwo has quit [Ping timeout: 250 seconds]
[00:04:22] theorb is now known as
theorbtwo
[00:12:16] -!-
acemi has quit [Quit: WeeChat 0.3.2]
[00:27:42] -!-
pjm__ has quit [Ping timeout: 250 seconds]
[00:33:17] -!-
DJUltis has quit [Remote host closed the connection]
[00:41:06] -!-
cjdavis1 has quit [Ping timeout: 240 seconds]
[00:54:17] -!-
andypugh has quit [Quit: andypugh]
[00:54:30] -!-
rooks has quit [Quit: So long, and thanks for all the fish.]
[00:58:22] -!-
i_tarzan has quit [Ping timeout: 252 seconds]
[01:04:46] <cradek> looks like I missed a lot of docs work - yay
[01:07:32] -!-
crazy_imp has quit [Ping timeout: 255 seconds]
[01:20:17] -!-
cjdavis has quit [Ping timeout: 276 seconds]
[01:27:46] <seb_kuzminsky> now master (and 2.5?) wont build on hardy because it build-depends on asciidoc >= 8.5, but 8.2.2 is the latest available on hardy
[01:28:01] <cradek> jepler made a pacakge I think
[01:28:16] <seb_kuzminsky> maybe just a buggy debian/control then
[01:29:04] <seb_kuzminsky> bbl
[01:29:06] -!-
seb_kuzminsky has quit [Quit: Client exiting]
[01:41:22] -!-
roberth__ has quit [Ping timeout: 250 seconds]
[01:42:35] -!-
i_tarzan has quit [Client Quit]
[01:58:23] -!-
krushia has quit [Ping timeout: 255 seconds]
[01:59:15] -!-
redwil has quit [Remote host closed the connection]
[02:02:52] <jepler> I've made a package on a personal machine
[02:02:55] <jepler> it's not online yet
[02:03:04] <jepler> I'll eventually put it on the linuxcnc.org package server
[02:03:41] -!-
adb has quit [Ping timeout: 260 seconds]
[02:04:07] -!-
VisualEcho has quit [Remote host closed the connection]
[02:05:01] <jepler> what?
[02:05:01] <jepler> $ apt-get source dblatex
[02:05:04] <jepler> E: Unable to find a source package for dblatex
[02:29:11] -!-
cpresser_ has quit [Ping timeout: 276 seconds]
[02:36:21] -!-
cpresser [
[email protected]] has joined #emc-devel
[02:42:38] <CIA-5> EMC: 03jepler 07v2.5_branch * r7e6c018c291b 10/ (debian/control.in src/configure.in): docs: check for working dblatex version
[02:42:47] <CIA-5> EMC: 03jepler 07v2.5_branch * r7a52656cf42d 10/src/configure.in: configure: fix text wrapping in error message
[02:42:47] <CIA-5> EMC: 03jepler 07v2.5_branch * rec3d0586c041 10/docs/src/config/ (ini_config.txt ini_config_fr.txt): Revert "docs: only use inline equations"
[02:42:50] <CIA-5> EMC: 03jepler 07v2.5_branch * r9104ebff187c 10/docs/src/common/user_intro.txt: Revert "docs: underscore in reference is bad"
[02:43:17] <jepler> situation is: pdf docs build on 8.04 but only with the versions of asciidoc and dblatex I've backported from lucid to hardy. Soon (tomorrow?) I'll get them on the linuxcnc.org package server. If I understand the buildbot's setup, it should automatically pick them up, but if not I'm sure seb can fix it.
[02:43:29] <jepler> I still haven't done a full package build, I dunno if there are other roadblocks.
[02:44:13] <jepler> 'night all
[03:43:46] -!-
mikeggg has quit [Ping timeout: 240 seconds]
[04:02:19] -!-
mhaberler [
[email protected]] has joined #emc-devel
[04:04:25] -!-
cpresser has quit [Ping timeout: 252 seconds]
[04:37:50] -!-
cpresser [
[email protected]] has joined #emc-devel
[04:38:52] -!-
Dannyboy has quit [Remote host closed the connection]
[04:45:51] -!-
cpresser has quit [Ping timeout: 260 seconds]
[04:45:56] -!-
cpresser [
[email protected]] has joined #emc-devel
[04:52:52] -!-
mhaberler has quit [Read error: Connection reset by peer]
[04:54:09] -!-
mhaberler [
[email protected]] has joined #emc-devel
[05:15:34] -!-
ve7it has quit [Remote host closed the connection]
[05:51:58] -!-
EDocTooR_ [
[email protected]] has joined #emc-devel
[05:55:18] -!-
EDocTooR has quit [Ping timeout: 250 seconds]
[06:16:16] -!-
cpresser has quit [Ping timeout: 260 seconds]
[06:21:56] -!-
cpresser [
[email protected]] has joined #emc-devel
[06:22:50] -!-
i_tarzan has quit [Ping timeout: 252 seconds]
[06:27:09] -!-
ca-kc6lbj-2 has quit [Remote host closed the connection]
[06:49:14] -!-
psha [
[email protected]] has joined #emc-devel
[07:18:30] <mhaberler> psha: hi! a question on source highlight - did you ever try it with gedit or just the gtksourceview widget? I'm getting odd coloring gedit'ing an ngc file.. (I'm trying to do a hal.lang def)
[07:33:27] <psha> no, i've just picked it, cleaned with soft cloth and put on it's place in emc2.git
[07:33:44] <psha> however it gives me not insane results on example gcodes
[08:03:29] -!-
cv has quit []
[08:05:56] -!-
aalpha1 has quit [Ping timeout: 260 seconds]
[08:11:13] -!-
Dannyboy has quit [Remote host closed the connection]
[08:38:15] -!-
mhaberler_ [
[email protected]] has joined #emc-devel
[08:40:27] -!-
mhaberler has quit [Ping timeout: 248 seconds]
[08:40:27] mhaberler_ is now known as
mhaberler
[09:06:40] <mhaberler> psha: here's an example of 'odd colouring':
http://imagebin.org/143306
[09:08:30] -!-
roberth__ [
[email protected]] has joined #emc-devel
[09:08:30] -!-
psha has quit [Read error: Connection reset by peer]
[09:08:35] -!-
mk0 has quit [Quit: Leaving]
[09:09:16] -!-
psha [
[email protected]] has joined #emc-devel
[09:09:47] <psha> and <extensions>
[09:09:58] <mhaberler> duh?
[09:10:26] <psha> 12:09 < psha> maybe it's not aware of comments?
[09:11:04] <mhaberler> my suspicicon is the gedit and gtksourcview versions of source-highlight arent compatible (v1.0 vs v2.0)
[09:12:27] <mhaberler> all I wanted is a decent source highlight for asciidocs.. maybe I can try that out without gedit
[09:12:37] <mhaberler> for hal files, that is
[09:22:57] <mhaberler> have you tried gcode formatting with it in asciidocs?
[09:28:29] <psha> hm, asciidoc uses either pygments or source-higlight packages
[09:28:38] <psha> not gtksourcesview highlight files
[10:01:35] -!-
PCW has quit [Remote host closed the connection]
[10:03:56] -!-
mikeggg has quit [Ping timeout: 264 seconds]
[10:23:51] <mhaberler> aha
[10:28:19] -!-
mhaberler has quit [Quit: mhaberler]
[11:02:30] -!-
pingufan has quit [Remote host closed the connection]
[11:05:36] -!-
Vq has quit [Ping timeout: 260 seconds]
[11:11:03] -!-
Guest905 has quit [Quit: Visitor from www.linuxcnc.org]
[11:59:44] -!-
skunkworks has quit [Ping timeout: 250 seconds]
[12:06:59] <jepler> psha: I see you copied a2x from 8.6.3. Besides the HTMLParser workaround, are you aware of any features in 8.6.3 but not in 8.5.2 that we require?
[12:10:41] -!-
mk0 has quit [Quit: Leaving]
[12:13:30] <psha> wait a bit
[12:14:07] -!-
rooks has quit [Quit: So long, and thanks for all the fish.]
[12:14:11] -!-
adb [
[email protected]] has joined #emc-devel
[12:15:36] <psha> nothing serious i see - some assorted fixes
[12:16:05] <psha> we require only htmllib related one
[12:18:28] <psha> good morning :)
[12:19:59] <jepler> hm, the buildbot doesn't live on irc anymore?
[12:20:04] <jepler> good morning
[12:21:18] <psha> what was it's nick? emc2-buildbot?
[12:21:54] <jepler> back in january it was going by emc2-buildmaster
[12:22:14] <jepler> emc-devel-2011-03-15.log:00:40 -!- emc2-buildmaster [
[email protected]] has quit [Ping timeout: 240 seconds]
[12:22:22] <jepler> (that's in my local time)
[12:22:39] <psha> it was here but silent
[12:22:56] <jepler> in principle I can /msg the bot to request that it try building
[12:22:57] <psha> so i can not check with log search :)
[12:23:12] <jepler> I put the asciidoc and dblatex packages on linuxcnc.org and wanted to see if it would pick them up..
[12:23:23] <jepler> it's not there for saying anything in the channel
[12:24:41] <jepler> well, this'll do it..
[12:24:47] <CIA-5> EMC: 03jepler 07master * rec3d0586c041 10/docs/src/config/ (ini_config.txt ini_config_fr.txt): Revert "docs: only use inline equations"
[12:24:48] <CIA-5> EMC: 03jepler 07master * r9104ebff187c 10/docs/src/common/user_intro.txt: Revert "docs: underscore in reference is bad"
[12:24:51] <CIA-5> EMC: 03jepler 07master * r7e6c018c291b 10/ (debian/control.in src/configure.in): docs: check for working dblatex version
[12:24:52] <CIA-5> EMC: 03jepler 07master * r7a52656cf42d 10/src/configure.in: configure: fix text wrapping in error message
[12:24:53] <CIA-5> EMC: 03jepler 07master * ra92aa176a0a6 10/ (5 files in 4 dirs): Merge remote branch 'origin/v2.5_branch'
[12:27:36] -!-
pingufan has quit [Remote host closed the connection]
[12:29:39] -!-
Jymmm has quit [Ping timeout: 240 seconds]
[12:29:39] -!-
redwil has quit [Ping timeout: 240 seconds]
[12:29:41] -!-
Sairon has quit [Ping timeout: 240 seconds]
[12:29:41] -!-
H264 has quit [Ping timeout: 240 seconds]
[12:29:41] -!-
MattyMatt has quit [Ping timeout: 240 seconds]
[12:29:43] -!-
Jymmmm [Jymmmm!~jymmm@unaffiliated/jymmm] has joined #emc-devel
[12:33:28] <jepler> hm, the buildbot didn't pick up the packages.
[12:43:07] -!-
skunkworks [skunkworks!447329d2@gateway/web/freenode/ip.68.115.41.210] has joined #emc-devel
[13:10:01] -!-
adb has quit [Ping timeout: 250 seconds]
[13:10:43] -!-
adb [
[email protected]] has joined #emc-devel
[13:20:09] -!-
Guest66940 has quit [Quit: Leaving]
[13:29:25] -!-
Athlocatle has quit [Quit: �����]
[13:53:38] <jepler> /-padx/com/edo
[13:53:41] <jepler> argh
[13:53:55] cjdavis1 is now known as
cjdavis
[13:54:19] <cradek> let's play "vi command or password?"
[14:12:10] -!-
mhaberler [
[email protected]] has joined #emc-devel
[14:14:32] <jepler> hm, looks like my ubuntu-hardy-server-amd64 vm is pretty sa
[14:14:34] <jepler> sad
[14:14:43] <jepler> /etc/rc2.d/S12dbus: 177: mkdir: not found
[14:14:59] <jepler> /sbin/lsmod: not found
[14:15:08] <cradek> what happened to it?
[14:15:12] <jepler> not sure
[14:15:39] <jepler> lots of messages about read-only filesystem, so it may be a damaged root fs
[14:16:36] <jepler> whatever's happened, important stuff like /bin/ls isn't there
[14:16:53] <psha> jepler: reboot with init=/bin/sh
[14:17:09] <jepler> I am able to reboot to the recovery shell
[14:17:13] <psha> it will boot in busybox which have everything built in
[14:17:26] <jepler> root@ubuntu-hrady-server-amd64:~# /bin/ls
[14:17:32] <jepler> bash: /bin/ls: No such file or directory
[14:17:54] <psha> plain 'ls' is not working?
[14:17:57] <jepler> no
[14:18:11] <jepler> there's no /bin/busybox for that matter
[14:18:34] <jepler> there's some stuff in bin (echo /bin/*) like bash, bunzip2, ..., zmore, znew
[14:18:43] <psha> that's bad
[14:19:47] <jepler> fsck says:
[14:19:58] <jepler> Entry 'ls' in /bin (755137) has deleted/unused inode 755186. Clear<y>?
[14:20:07] <jepler> too bad, I think the filesystem is toast ..
[14:21:48] <skunkworks> hard drive issue? any smart data?
[14:22:09] <jepler> it's a vm. the host's drive is fine as far as I can discern.
[14:22:16] <skunkworks> oh - vm
[14:23:42] <jepler> yeah
[14:26:15] -!-
izua has quit [Ping timeout: 276 seconds]
[14:40:30] -!-
aalpha1 has quit [Quit: �I-n-v-i-s-i-o-n� 2.0 Build 2520]
[14:41:30] -!-
crazy_imp has quit [Ping timeout: 241 seconds]
[14:47:57] -!-
adb has quit [Ping timeout: 250 seconds]
[14:51:51] -!-
seb_kuzminsky [
[email protected]] has joined #emc-devel
[15:08:49] -!-
adb [
[email protected]] has joined #emc-devel
[15:32:36] -!-
awallin_ [awallin_!~quassel@2001:708:110:1020:224:7eff:feda:7c7d] has joined #emc-devel
[15:35:08] -!-
adb has quit [Ping timeout: 264 seconds]
[15:35:33] -!-
nullie has quit [Quit: Ex-Chat]
[15:36:52] <jepler> seb_kuzminsky: I uploaded new asciidoc and dblatex packages to the linuxcnc.org package server, but the buildbot didn't pick them up. Any idea why? How can I kick it to make it try again?
[15:38:38] <seb_kuzminsky> new .debs for hardy?
[15:38:47] <jepler> right
[15:38:52] <seb_kuzminsky> i'll update
[15:39:02] -!-
ve7it [
[email protected]] has joined #emc-devel
[15:39:29] <jepler> http://pastebin.com/a9U2vutJ
[15:44:47] <seb_kuzminsky> upgraded, rebuilding now
[15:46:15] <jepler> thank you
[15:46:25] <jepler> is it because it doesn't 'apt-get update' automatically?
[15:47:01] <seb_kuzminsky> it's my crappy script that tries and fails to install missing build dependencies
[15:47:13] <seb_kuzminsky> http://buildbot.linuxcnc.org/buildbot/builders/hardy-i386-trunkish-realtime-rip/builds/186/steps/install-missing-build-dependencies/logs/stdio
[15:47:31] <seb_kuzminsky> it runs dpkg-checkbuilddeps, in runs "apt-get install" on anything it complains about
[15:47:39] <seb_kuzminsky> but it doesn't handle versioned dependencies :-(
[15:47:45] <jepler> oh
[15:47:46] <jepler> I see
[15:48:13] <seb_kuzminsky> when building the deb inside pbuilder later it should all work fine, because pbuilder does that stuff correctly
[15:48:27] <seb_kuzminsky> thanks for fixing the docs
[15:49:47] <jepler> maybe this is something you can steal?
http://www.linux.codehelp.co.uk/serendipity/index.php?/archives/222-guid.html
[15:50:04] <seb_kuzminsky> heh
[15:50:17] -!-
mhaberler has quit [Quit: mhaberler]
[15:51:10] <cradek> seb_kuzminsky: do you think you can still have buildbot do the work of building releases for me, based on a signed commit or whatever we figured out?
[15:51:52] <jepler> hm, the script doesn't work for me out-of-the-box. :(
http://pastebin.com/9KUvZpAv
[15:52:35] <seb_kuzminsky> cradek: yes i think so
[15:52:59] <cradek> whee
[15:53:02] <cradek> that makes me happy
[15:53:14] <seb_kuzminsky> i think if you push a commit that git-describes to a simple version number, it'll just automatically build it and it'll look like a real release
[15:54:06] <seb_kuzminsky> you can try it by making a "2.5.0-pre2" practice release and we'll see if it does it right
[15:55:26] <seb_kuzminsky> i think the sequence is something like this:
[15:55:44] <seb_kuzminsky> update debian/changelog and VERSION, commit, tag, push
[16:07:26] -!-
psha[note] [psha[note]
[email protected]] has joined #emc-devel
[16:07:45] <cradek> cool, I will try it ... but not yet
[16:08:19] <cradek> it's really coming together - thanks for everyone's work
[16:08:36] <cradek> I hope to spend some more time on updating docs.
[16:08:42] -!-
mhaberler [
[email protected]] has joined #emc-devel
[16:09:09] -!-
mhaberler_ [
[email protected]] has joined #emc-devel
[16:24:13] <seb_kuzminsky> hm, the hardy sim pbuilders are missing the linuxcnc deb archive, i'll fix that
[16:24:48] -!-
ca-kc6lbj-2 has quit [Remote host closed the connection]
[16:28:01] -!-
mk0 has quit [Quit: Leaving]
[16:28:07] -!-
psha has quit [Quit: Lost terminal]
[16:28:52] <jepler> but the deb-hardy-rt-binary packages built again for the first time since the doc merge -- woo
[16:29:07] <cradek> yayyy!
[16:29:09] <cradek> thanks jepler
[16:29:24] <jepler> my pleasure
[16:34:17] -!-
skunkworks has quit [Ping timeout: 252 seconds]
[16:54:08] -!-
mhaberler_ has quit [Quit: mhaberler_]
[17:11:11] -!-
nullie has quit [Read error: Operation timed out]
[17:46:23] <seb_kuzminsky> ooh, it's looking green!
[18:01:32] -!-
awallin_ has quit [Remote host closed the connection]
[18:15:31] -!-
adb [
[email protected]] has joined #emc-devel
[18:21:23] <jepler> psha[note]: it's not a bad thing that I had to build asciidoc and dblatex to make the doc toolchain work on hardy
[18:21:58] <jepler> it's a good thing that it was feasible to. The one time I did try it (6.06 / 8.04 days), I was unable to build the debian packages of the same LyX version on both systems..
[18:22:10] <jepler> leading to one of the major problems with the old doc toolchain
[18:26:31] -!-
mhaberler has quit [Quit: mhaberler]
[18:28:22] <jepler> whee I see that the build is now successful across the board
[18:33:06] <psha[note]> jepler: glad to hear that :) otherwise only way for me was to make seppuku :)
[18:43:39] -!-
grandrew_ has quit [Ping timeout: 276 seconds]
[18:49:30] -!-
grandrew__ has quit [Ping timeout: 276 seconds]
[18:53:10] -!-
psha[note] has quit [Quit: Lost terminal]
[18:54:48] -!-
psha [
[email protected]] has joined #emc-devel
[18:58:18] -!-
andypugh [andypugh!~andy2@cpc2-basl1-0-0-cust1037.basl.cable.virginmedia.com] has joined #emc-devel
[19:02:50] -!-
grandrew_ has quit [Read error: Connection reset by peer]
[19:06:17] -!-
steves_logging has quit [Ping timeout: 276 seconds]
[19:11:05] <cradek> someone with methodical troubleshooting skills who can reproduce this stepgen problem could use git bisect to find the commit that breaks it
[19:12:00] <cradek> seems like it has morphed from a bug "on 10.04" into "after 2.4.1 sometime when running on 8.04"
[19:12:44] <cradek> the bug is reported on 2.6.0, which doesn't exist, and I haven't seen anyone say they've made sure the configs match
[19:13:57] <cradek> ah, cmorley is on it now
[19:14:14] <cradek> maybe he can test methodically and bisect it, if it's an emc (not config) bug
[19:21:14] <cradek> two commits affected stepgen.c between 2.4.6 and master: 721715f8eda5 and 1f39204a0722
[19:22:34] <cradek> well ... there are a couple others
[19:23:22] <jepler> between 2.4.6 and master? I get a different list than you. $ git log --oneline v2.4.6..origin/master -- src/hal/components/stepgen.c
[19:23:26] <jepler> 3b250c9 Merge remote branch 'origin/v2.4_branch'
[19:23:28] <jepler> 733af60 Allow the user to specify a stepping waveform
[19:23:46] <cradek> src/hal/drivers/mesa-hostmot2/stepgen.c
[19:23:55] <jepler> durrrrr
[19:24:38] <cradek> don't be too hard on yourself - I think it has been claimed that software stepgen is also broken
[19:27:25] <cradek> "and this configuration is a software stepgen and do the same (joint folloving error)."
[19:27:53] <psha> cradek: 2.6.0 means that this bug will be introduced somewhere this year?
[19:28:11] <psha> or just git master introduces it as 2.6?
[19:28:21] <cradek> 2.6.0 does not exist so I don't know what a user means when he says the bug is in 2.6.0
[19:28:26] <seb_kuzminsky> master calls itself 2.6.0-pre0
[19:29:07] <cradek> seb_kuzminsky: have you tried to reproduce this?
[19:29:27] <seb_kuzminsky> link?
[19:29:34] <seb_kuzminsky> (ie: not yet)
[19:30:12] <cradek> afaic you will get no benefit from reading the thread, but:
http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/id,7681/catid,10/limit,6/limitstart,0/lang,english/
[19:31:21] <cradek> sounds like cmorley has seen unexpected following errors, but he hasn't tried to find the problem
[19:46:23] -!-
adb has quit [Ping timeout: 252 seconds]
[19:53:00] -!-
mhaberler [
[email protected]] has joined #emc-devel
[19:56:10] <mhaberler> psha: did you ever try to format an asciidoc source,ngc block?
[19:58:00] <psha> no
[19:58:38] <psha> but all you need is to teach either source-highlight or pygments about ngc
[19:58:52] <psha> asciidoc passes language name to filter
[20:06:23] -!-
adb [
[email protected]] has joined #emc-devel
[20:38:38] davec_ is now known as
Guest15728
[20:40:48] -!-
cevad has quit [Ping timeout: 240 seconds]
[21:09:43] -!-
mk0 has quit [Ping timeout: 250 seconds]
[21:17:25] -!-
motioncontrol has quit [Quit: Sto andando via]
[21:20:16] -!-
x_ [
[email protected]] has joined #emc-devel
[21:36:39] -!-
psha has quit [Quit: zzz]
[21:52:15] -!-
pjm has quit [Ping timeout: 246 seconds]
[21:52:54] x_ is now known as
micges
[21:54:52] -!-
mozmck has quit [Quit: Leaving.]
[22:05:31] -!-
mozmck [mozmck!~moses@client-173.225.233.219.dfwtx.partnershipbroadband.com] has joined #emc-devel
[22:06:50] -!-
adb has quit [Ping timeout: 250 seconds]
[22:07:05] <andypugh> Is it enough to have a #define of a header file with a function prototype in it for the file to find a function?
[22:07:24] <andypugh> Or do I need to do something cleverer with externs and such?
[22:08:55] <alex_joni> do you mean #include ?
[22:09:21] <andypugh> Err, yes.
[22:09:26] <andypugh> (Checks)
[22:09:37] <andypugh> Yes.
[22:09:53] <alex_joni> you usually include the .h, where the function prototype is defined
[22:09:58] <alex_joni> then the compiling works
[22:10:12] <andypugh> I am getting the message "hm2: Unknown symbol hm2_get_bspi"
[22:10:17] <alex_joni> but for linking you also need the .c (more specific the .o)
[22:10:21] <andypugh> It all compiles...
[22:10:56] <andypugh> hm2_get_bspi is a function in another file.
[22:12:05] <alex_joni> if it's not immediate obvious where the function is, you can name it extern, then the linker won't check so closely
[22:12:14] <alex_joni> just make sure it's available when you call it
[22:12:34] <andypugh> I get the error when I call it. There are no compiler warnings.
[22:13:39] <andypugh> A complication is that the function is being called from a comp. I might need to look at the generated C-code
[22:14:51] -!-
alex_chally has quit [Quit: ZNC -
http://znc.sourceforge.net]
[22:16:20] -!-
rooks has quit [Read error: Operation timed out]
[22:17:35] <andypugh> http://pastebin.com/wxjPHMGj
[22:18:27] <andypugh> hostmot.h contains the function prototype. The #include is in a slightly odd place because of the way the comp preprocessor does stuff.
[22:18:54] <andypugh> I _don't_ see the result of the HM2_PRINT statement before the call to hm2_get_bspi
[22:19:08] <alex_joni> hmm.. so how do you compile it?
[22:19:16] <andypugh> make
[22:19:27] <alex_joni> makefile?
[22:19:39] <andypugh> The normal one
[22:20:08] <alex_joni> and you added your source file?
[22:20:19] <andypugh> The comp is in the hal/components directory so gets picked up by the generic submakefile
[22:20:50] <andypugh> The get_count function runs when I addf the module.
[22:21:12] <alex_joni> getting picked up doesn't mean it gets the proper .o linked in
[22:21:20] <alex_joni> I would suspect..
[22:21:51] <andypugh> I have made lots of other comps, and they have all worked. None have called external functions though.
[22:22:07] <alex_joni> that's my point
[22:22:21] <alex_joni> I think doing this as a comp is overcomplicating things
[22:22:40] <andypugh> I am trying to prove that it can be done.
[22:23:22] <andypugh> The idea is to make making other sub-drivers for other SPI devices relatively easy
[22:23:46] <andypugh> 'course a proper programmer would be a good starting point
[22:24:33] <alex_joni> I doubt a comp would be of an advantage in this case
[22:24:58] <alex_joni> I mean who attempts to write a subdriver for hostmot surely can cope with 30 extra lines of c-code
[22:25:08] <alex_joni> (which are already there from your example btw)
[22:27:26] -!-
dgarr [
[email protected]] has joined #emc-devel
[22:27:50] <andypugh> The pin definitions and handling of multiple instances is a lot tidier in a comp.
[22:28:22] <andypugh> self-documenting too.
[22:28:47] <alex_joni> then do that, rename the generated file and pickup from there :)
[22:29:23] <andypugh> Any idea how to add a comp to the makefile?
[22:30:02] <alex_joni> maybe put it someplace else, and use a similar incantation for generating the .c file, then a regular compiling/linking
[22:30:44] -!-
isssy has quit [Quit: Visitor from www.linuxcnc.org]
[22:31:03] <cpresser> alex_joni: comp has a option "--preprocess"
[22:31:23] <cpresser> use that and make sure it is run bevore CC
[22:32:21] <andypugh> I have absolutely no idea where makefiles are concerned
[22:32:37] <seb_kuzminsky> andypugh: look at how hostmot2 exposes hm2_register() to the low-level anyio drivers
[22:35:12] <seb_kuzminsky> EXPORT_SYMBOL_GPL(hm2_register);
[22:35:15] <seb_kuzminsky> bbl
[22:35:36] <andypugh> it just appears to be declared in hm2-lowlevel.h, which is included in hostmot2.h, or am I missing something?
[22:35:53] <andypugh> Looks like I am missing something :-)
[22:37:52] <cpresser> perhaps i dont see the whole picture, but may it be possible that you have to link you file to the one conatining the function hm2_get_bspi
[22:38:21] <micges> andypugh: pastebin source, i'll try it
[22:38:40] <cpresser> ah no, take that back.. you said that tge file compiles, its a runtime error
[22:38:56] <andypugh> I think seb has it with the EXPORT_SYMBOL
[22:41:12] <andypugh> w00t! [ 6154.265436] BUG: unable to handle kernel NULL pointer dereference at 00000014
[22:41:13] <andypugh> [ 6154.265447] IP: [<fa5e25e0>] rtapi_vsnprintf+0x2d9/0x6aa [rtapi]
[22:41:28] <andypugh> Never thought I would be glad to see that error,:-)
[22:46:07] -!-
micges has quit [Quit: Ex-Chat]
[22:50:27] -!-
renesis has quit [Read error: Connection reset by peer]
[22:53:54] -!-
Fox_Muldr has quit [Ping timeout: 276 seconds]
[23:06:14] -!-
Ikarus has quit [Read error: Operation timed out]
[23:09:42] -!-
tlab has quit [Quit: Leaving]
[23:10:24] <jepler> for one hal component to "see" a function defined in another hal component, the component defining the function must (A) export the function with EXPORT_SYMBOL_GPL or EXPORT_SYMBOL and (B) be loaded before the component using the function.
[23:11:04] <jepler> for instance, here's one of the existing EXPORT_SYMBOL_GPL from hostmot2: hostmot2.c:EXPORT_SYMBOL_GPL(hm2_register);
[23:11:31] <jepler> this is the function that lets an individual card driver be registered
[23:11:45] <andypugh> I have them all talking to each other now. It's crashing RTAI every time, but you can't have everything. I am totally baffled by strings in C
[23:12:17] <jepler> HM2_PRINT_NO_LL is just a macro that expands to rtapi_print, which is already exported.
[23:12:48] <andypugh> Yes, I am using it to trace program flow.
[23:13:07] <andypugh> (because I have not invested the time to learn a debugger
[23:13:39] <jepler> afaik there's only an effective debugger for components that can be built with --enable-simulator, so you'd still be flying blind working on hostmot2
[23:13:45] <andypugh> I am getting very fast at typing "sudo shutdown -r now"
[23:14:49] <jepler> if you want hm2_get_bspi to change the value of b inside EXTRA_SETUP (e.g., to a new hm2_bspi_instance that hm2_get_bspi just allocated), then you have to pass &b to it, and the corresponding item in the prototype would be hm2_bspi_instance**
[23:15:34] <jepler> otherwise the pointer b in EXTRA_SETUP is never initialized before you dereference it at b->name
[23:15:36] <andypugh> I thought I had defined b as a pointer?
[23:16:13] <andypugh> I think this is one of those things that you either get straight away, or end up forever second-guessign yourself on.
[23:16:40] <jepler> Yes, the type of b is pointer to hm2_bspi_instance. But the function of hm2_bspi_instance is to change what b points at (the value of b). In C (and unlike in some other languages), when you want to change the value of something passed in, you do it by passing the location of the thing to be modified
[23:16:52] <jepler> that means you pass in &b, and that means the type of the parameter gets more stars
[23:16:57] <andypugh> Ah, yes.
[23:17:07] <andypugh> Pass-by-value catches me out again
[23:17:24] <andypugh> (I have spent years in VBA, which is always pass-by-reference)
[23:17:34] <jepler> yes, it's often difficult to shift gears like that
[23:18:14] <andypugh> VBA is OO too, so I am forever swapping object identities with each other.
[23:19:17] <andypugh> This is probably sacrelige, but I rather like VBA and the IDE in Excel. It has it's quirks, though.
[23:19:59] <andypugh> So: int hm2_get_bspi(hm2_bspi_instance_t **q, char *name){
[23:20:07] <andypugh> q = &all_bspis[i]->instance[j];
[23:20:26] <andypugh> I am guessing that that q needs a layer of indirection?
[23:20:42] -!-
Dallur1 has quit [Quit: Leaving.]
[23:36:01] -!-
acemi has quit [Quit: WeeChat 0.3.2]